Testing coverage based reliability modeling for multi-release open-source software incorporating fault reduction factor

Abstract

Testing represents an important and necessary phase in the development process of a software that provides a reliable and efficient software product. To enhance software reliability, fault detection and correction are carried out during this process. In order to develop highly reliable software system software reliability growth models (SRGMs) are observed to be the most effective way. These SRGMs have been developed under a set of assumptions, there are many real-life factors that control the reliability of the software system during the testing process. In that context Fault Reduction Factor (FRF) and testing coverage are the most important key factors. In our proposed study we have developed SRGMs based on these environmental factors. We have also discussed the concept of multi releases of an Open-Source Software (OSS) project as a single release of the product may not meet the ever-increasing requirements of a competitive market. Therefore, this paper talks about the impact of different parameters such as FRF and testing coverage on the reliability of the different releases of an OSS. The mathematical expression based on these assumptions is provided by two models. These models assume the FRF as constant and testing coverage is assumed to have Weibull and Exponentiated Weibull distribution functions. The proposed models are then validated on real-life faults datasets and their performance is measured on the basis of different criteria of the goodness of fit.